CN103557841A - Method for improving photogrammetric precision of multi-camera resultant image - Google Patents

Abstract

The invention discloses a method for improving the photogrammetric precision of a multi-camera resultant image. The method comprises the following steps: accurately converting image point coordinates directly observed in the photogrammetric treatments comprising the aerial triangulation adjustment, three-dimensional measurement and ortho-rectification of the ortho-rectification onto single-camera raw images to obtain image points on the single-camera raw images, combining the exterior orientation elements of the multi-camera resultant image with the orientation elements of single cameras, determined in a multi-camera system to make the image points on the single-camera raw images, and single-camera object lens centers and corresponding ground points thereof strictly satisfy photogrammetric collinear geometry conditions, and carrying out rigorous photogrammetric treatment in order to improve the photogrammetric precision of a multi-camera resultant image.

Description

A kind of method that improves polyphaser resultant image photogrammetric accuracy

Technical field

The present invention relates to photogrammetric field, be specifically related to a kind of method that improves polyphaser resultant image photogrammetric accuracy.

Background technology

Photogrammetric is the technology that the image of the space object taken by camera measures the shape of space object, size, locus, character and mutual relationship.In order to improve the precision of measurement and the efficiency of measurement, modern photography mostly many one cameras is equivalent to a wide angle camera by different photography directions together with position grouping in measuring or panorama camera is photographed to the diverse location of space object and direction simultaneously, as far as possible once obtain to wide-angle the image of space object on a large scale, through subsequent calculations machine, processing the image joint that these one cameras are taken again becomes one to be equivalent to one and to only have the virtual wide angle camera of an equivalent projection centre or the resultant image that panorama camera is taken, then resultant image is carried out Photogrammetric Processing and reaches measurement object, also the resultant image that can reinstate such splicing with Other Instruments equipment one positions navigation purposes.Because such polyphaser resultant image has certain projection approximation in splicing, not the central projection image of stricti jurise, can only just can reach the accuracy requirement of measurement meeting under certain photography conditions to use.

Document 1 (Liang Tang, Christoph karsten Jacobsen, Christian Heipke, Alexander Hinz, 2000.GEOMETRIC ACCURACY POTENTIAL OF THE DIGITAL MODULAR CAMERA, International Archives of Photogrammetry and Remote Sensing.Vol.XXXIII, Part B4.Amsterdam2000, pp.1051～1057.) photogrammetric accuracy of aviation face battle array polyphaser resultant image is analyzed, and provided the relation curve of photogrammetric accuracy between photo distance.

Document 2 (Cramer, M., 2007.The EuroSDR Performance Test for Digital Aerial Camera Systems.Proceedings of the 51st Photogrammetric Week, Stuttgart, Germany.3-7September, pp.89-106.), document 3 (Madani M., Shkolnikov is GEOMETRIC ACCURACY OF DMC ' S VIRTUAL IMAGES.TheInternational Archives of the Photogrammetry I.2008.INCREASING, Remote Sensing and Spatial Information Sciences.Vol.XXXVII.Part B3a.2008Beijing.) all pointed out that polyphaser resultant image exists the problem that measurement accuracy is low, and proposed by empty three adjustments of the self calibration in photogrammetric and tested the methods such as rear error graticule mesh correction and improve precision.

At present, because multicamera system kind is numerous, performance difference is larger, and the complexity of photography conditions is various, and empty three method of adjustment of the self calibration of above-mentioned document proposition are difficult to select suitable additional correction parameter to obtain stable measurement, test the rear error graticule mesh correction method systematic error that also difficulty exists resultant image in the ground control point situation of some and carry out effective compensation.Especially when close-range photography, these methods all do not reach the accuracy requirement of measurement, can not directly to the performance of each one camera in multicamera system, carry out regular calibration.And the accurate measurement of the multicamera system of using for ground close-range photography also lacks similar example at present, lack the ways and means that all polyphaser resultant images are carried out to effectively tight processing that versatility is stronger.

Summary of the invention

Problem in view of the existence of existing photogrammetric technology field, the present invention is only to propose a kind of method that improves polyphaser resultant image photogrammetric accuracy, mainly solves the low problem of photogrammetric accuracy while existing projection approximation to cause close-range photography in polyphaser resultant image splicing.

Say further, the present invention utilizes Conversion Relations tight between polyphaser resultant image when splicing and one camera raw video, by polyphaser resultant image in empty three adjustments, the picpointed coordinate of observing in the Photogrammetric Processing such as three-dimensional measuring and orthorectify accurately transforms to the picpointed coordinate that obtains correspondence on one camera raw video on one camera raw video, again the elements of exterior orientation of polyphaser resultant image and one camera definite orientation element in multicamera system is combined, set up the corresponding picture point on one camera raw video, tight photogrammetric collinearity equation between ground point corresponding to one camera object lens center picture point corresponding to this under this correspondence picture point, it is carried out tight Photogrammetric Processing and reaches the object that improves measurement accuracy.

The present invention is theoretical tight, and highly versatile, not only can process the polyphaser resultant image in low latitude photography situation, also can process the polyphaser resultant image that ground close-range photography obtains, and comprises full-view image.

To achieve these goals, the technical solution used in the present invention is as follows:

A method that improves polyphaser resultant image photogrammetric accuracy, said method comprising the steps of:

(1) multicamera system is carried out to calibration obtains the distortion of each one camera elements of interior orientation in multicamera system, object lens and the camera parameter of definite elements of exterior orientation in polyphaser coordinate system;

(2) described multicamera system is installed on camera carrying platform according to the requirement of measure the item and supporting positioning navigating device is photographed to the target in described measure the item location, and is obtained raw video and the location navigation data of all one cameras in described multicamera system;

(3) according to the requirement of measure the item, the ground control point in project location district is observed; (4) according to the selected polyphaser resultant image of the type of described multicamera system, splice required Method of Projection Change, and the camera parameter obtaining according to step (1) is set up image coordinate Conversion Relations tight between one camera raw video and polyphaser resultant image;

(5) each one camera raw video step (2) being obtained, then the image coordinate Conversion Relations obtaining by step (4) splices, obtain each polyphaser resultant image and elements of interior orientation thereof, and on polyphaser resultant image, mark the corresponding region of one camera raw video;

(6), according to the requirement of measure the item, the one camera parameter of utilizing location navigation data that step (2) obtains, polyphaser resultant image that step (5) obtains and elements of interior orientation thereof and step (1) to obtain is set up the photogrammetric data that empty three adjustments, three-dimensional measuring and orthorectify use and is processed engineering;

(7) photogrammetric data of setting up according to step (6) process polyphaser resultant image that engineering obtains step (5) be encrypted a picpointed coordinate (what is manual observation with manual observation automatically?), the corresponding picpointed coordinate of the ground control point that step (3) is obtained on polyphaser resultant image carries out manual observation, and the corresponding region of one camera raw video marking on the polyphaser resultant image that the image coordinate of this 2 class point on polyphaser resultant image obtained by the image coordinate Conversion Relations of step (4) foundation and step (5) transforms on one camera raw video, obtain the image coordinate on one camera raw video,

(8) picture point on the one camera raw video that the location navigation data that integrating step (2) obtains obtain step (7), the described picture point place one camera raw video projection centre ground point corresponding with it are set up respectively corresponding collinearity equation formula according to the pattern of aeroplane photography or ground photography, and the collinearity equation formula under aeroplane photography pattern is:

$\left[\begin{array}{c}X\\ Y\\ Z\end{array}\right]=\left[\begin{array}{c}{X}_{\mathrm{GPSO}}+X_{\mathrm{GPS}}\\ Y_{\mathrm{GPSO}}+Y_{\mathrm{GPS}}\\ Z_{\mathrm{GPSO}}+Z_{\mathrm{GPS}}\end{array}\right]+R_{\mathrm{IMUO}}\·R_{\mathrm{IMU}}\·R_{\mathrm{virt}}(\left[\begin{array}{c}\mathrm{\Δ}{X}_{\mathrm{GPS}-\mathrm{virt}}\\ \mathrm{\Δ}{Y}_{\mathrm{GPS}-\mathrm{virt}}\\ \mathrm{\Δ}{Z}_{\mathrm{GPS}-\mathrm{virt}}\end{array}\right]+\left[\begin{array}{c}{X}_S^{\mathrm{sub}}\\ Y_S^{\mathrm{sub}}\\ Z_S^{\mathrm{sub}}\end{array}\right]+\mathrm{\λ}{M}_{\mathrm{sub}}\left[\begin{array}{c}{x}_{\mathrm{sub}}-x_{\mathrm{sub}}^0+\mathrm{\Δ}{x}_{\mathrm{sub}}\\ y_{\mathrm{sub}}-y_{\mathrm{sub}}^0+\mathrm{\Δ}{y}_{\mathrm{sub}}\\ -f_{\mathrm{sub}}\end{array}\right])---\left(1\right)$

Wherein, X, Y, Z is the three-dimensional coordinate of millet cake in object space mapping coordinate system accordingly, X _gPS, Y _gPS, Z _gPSfor the three-dimensional coordinate of taking the photograph station that positioning navigating device in step (2) is obtained in object space mapping coordinate system, X _gPSO, Y _gPSO, Z _gPSOfor X _gPS, Y _gPS, Z _gPSthe value of starting in object space mapping coordinate system, R _iMUfor the attitude angle of taking the photograph station that positioning navigating device in step 2 is obtained in object space mapping coordinate system, the rotation matrix being formed by roll angle ω, angle of pitch φ and position angle κ, R _iMUOfor the attitude angle of taking the photograph station value of starting in object space mapping coordinate system, the rotation matrix being formed by roll angle ω 0, angle of pitch φ 0 and position angle κ 0, R _virtfor the inconsistent amount of positioning navigating device coordinate system direction in multicamera system coordinate system and step (1), the rotation matrix being formed by roll angle ω virt, angle of pitch φ virt and position angle κ virt, M _subfor one camera definite deflection element in multicamera system coordinate system, the rotation matrix being formed by roll angle ω sub, angle of pitch φ sub and position angle κ sub, Δ X _gPS-virt, Δ Y _gPS-virt, Δ Z _gPS-virtfor gps antenna center definite relative position in polyphaser coordinate system,

for one camera object lens center definite relative position in polyphaser coordinate system, x _sub, y _subpicture point image coordinate on the one camera raw video obtaining for step (7),

for the principal point position of one camera, f _subfor the main distance of one camera, Δ x _sub, Δ y _subfor x _sub, y _subin the systematic error compensation that caused by one camera deformation of image and object lens distortion etc.;

Collinearity equation formula under ground photography pattern is:

$\left[\begin{array}{c}X\\ Y\\ Z\end{array}\right]=\left[\begin{array}{c}{X}_{\mathrm{GPSO}}+X_{\mathrm{GPS}}\\ Y_{\mathrm{GPSO}}+Y_{\mathrm{GPS}}\\ Z_{\mathrm{GPSO}}+Z_{\mathrm{GPS}}\end{array}\right]+R_{\mathrm{IMUO}}\·R_{\mathrm{IMU}}\·R_{\mathrm{virt}}(\left[\begin{array}{c}\mathrm{\Δ}{X}_{\mathrm{GPS}-\mathrm{virt}}\\ \mathrm{\Δ}{Y}_{\mathrm{GPS}-\mathrm{virt}}\\ \mathrm{\Δ}{Z}_{\mathrm{GPS}-\mathrm{virt}}\end{array}\right]+\left[\begin{array}{c}{X}_S^{\mathrm{sub}}\\ Y_S^{\mathrm{sub}}\\ Z_S^{\mathrm{sub}}\end{array}\right]+\mathrm{\λ}{M}_{\mathrm{sub}}\left[\begin{array}{c}{x}_{\mathrm{sub}}-x_{\mathrm{sub}}^0+\mathrm{\Δ}{x}_{\mathrm{sub}}\\ f_{\mathrm{sub}}\\ y_{\mathrm{sub}}-y_{\mathrm{sub}}^0+\mathrm{\Δ}{y}_{\mathrm{sub}}\end{array}\right])---\left(2\right)$

Wherein, R _iMUO, R _iMU, R _virtand M _subbe respectively the rotation matrix usually forming according to the angle unit of ground photograph mode definition, the implication of other symbol representative is identical with formula (1);

(9) the aeroplane photography pattern of step (8) being set up and the collinearity equation formula under ground photography pattern are all carried out linearization by same mode and are obtained following identic error equation:

$\left\{\begin{array}{cccccccc}{V}_P=C_{\mathrm{GPSIMU}}{X}_{\mathrm{GPSIMU}}+C_G{X}_G+C_A{X}_A+C_{\mathrm{GPS}-\mathrm{virt}}{X}_{\mathrm{GPS}-\mathrm{virt}}+C_{\mathrm{IMU}-\mathrm{virt}}{X}_{\mathrm{IMU}-\mathrm{virt}}+C_{\mathrm{virt}-\mathrm{sub}}{X}_{\mathrm{virt}-\mathrm{sub}}& & & & & & -L_P& P_P\\ V_G=& E_G{X}_G& & & & & -L_G& P_G\\ V_A=& & E_A{X}_A& & & & -L_A& P_A\\ V_{\mathrm{GPSIMU}}=E_{\mathrm{GPSIMU}}{X}_{\mathrm{GPSIMU}}& & & & & & -L_{\mathrm{GPSIMU}}& P_{\mathrm{GPSIMU}}\\ V_{\mathrm{GPS}-\mathrm{virt}}=& & & E_{\mathrm{GPS}-\mathrm{virt}}{X}_{\mathrm{GPS}-\mathrm{virt}}& & & -L_{\mathrm{GPS}-\mathrm{virt}}& P_{\mathrm{GPS}-\mathrm{virt}}\\ V_{\mathrm{IMU}-\mathrm{virt}}=& & & & E_{\mathrm{IMU}-\mathrm{virt}}{X}_{\mathrm{IMU}-\mathrm{virt}}& & -L_{\mathrm{IMU}-\mathrm{virt}}& P_{\mathrm{IMU}-\mathrm{virt}}\\ V_{\mathrm{virt}-\mathrm{sub}}=& & & & & E_{\mathrm{virt}-\mathrm{sub}}{X}_{\mathrm{virt}-\mathrm{sub}}& -L_{\mathrm{virt}-\mathrm{sub}}& P_{\mathrm{virt}-\mathrm{sub}}\end{array}\right.---\left(3\right)$

Wherein, VP, VG, VA, VGPS-IMU, VGPS-virt, VIMU-virt and Vvirt-sub are respectively one camera raw video coordinate, ground point object space three-dimensional coordinate, self calibration additional parameter, GPS/IMU location navigation data, gps antenna center is to polyphaser resultant image projection centre side-play amount, polyphaser coordinate system and IMU coordinate system are inconsistent, and the observed reading of the orientation element of one camera in polyphaser coordinate system corrects vector, XG, XA, XGPS-IMU, XGPS-virt, XIMU-virt and Xvirt-sub are respectively ground point object space three-dimensional coordinate, self calibration additional parameter, GPS/IMU location navigation data, gps antenna center is to polyphaser resultant image projection centre side-play amount, polyphaser coordinate system and IMU coordinate system are inconsistent, and the orientation element of one camera in polyphaser coordinate system is as the correction vector of adjustment unknown number, CG, CA, CGPS-IMU, CGPS-virt, CIMU-virt and Cvirt-sub are respectively XG, XA, XGPS-IMU, XGPS-virt, the matrix of coefficients that XIMU-virt and Xvirt-sub are corresponding, EG, EA, EGPS-IMU, EGPS-virt, EIMU-virt and Evirt-sub are respectively XG, XA, XGPS-IMU, XGPS-virt, the unit coefficient matrix that XIMU-virt and Xvirt-sub are corresponding, LP, LG, LA, LGPS-IMU, LGPS-virt, LIMU-virt and Lvirt-sub are respectively VP, VG, VA, VGPS-IMU, VGPS-virt, the constant vector that VIMU-virt and Vvirt-sub are corresponding, PP, PG, PA, PGPS-IMU, PGPS-virt, PIMU-virt and Pvirt-sub are respectively one camera raw video coordinate, ground point object space three-dimensional coordinate, self calibration additional parameter, GPS/IMU location navigation data, gps antenna center is to polyphaser resultant image projection centre side-play amount, polyphaser coordinate system and IMU coordinate system are inconsistent, and the weight matrix corresponding to observed reading of the orientation element of one camera in polyphaser coordinate system, the precision that reflects these observed readings,

(10) according to the method for photogrammetric hollow three adjustments, by the error equation that least square method is set up step (9) methodization of carrying out, form normal equation, by interative computation, try to achieve the unknown number in error equation, when resolving sky three adjustment processing that complete polyphaser resultant image while meeting predefined precision conditions, obtain the end product of empty three adjustments;

(11) according to sky three adjustment result that obtain, polyphaser resultant image is carried out to three-dimensional measuring, the picpointed coordinate of observing on polyphaser resultant image is transformed on one camera raw video by processing mode step (7) Suo Shu, obtain the image coordinate on one camera raw video, again by the orientation element combination in multicamera system with corresponding one camera of the elements of exterior orientation of the multicamera system in the sky obtaining three adjustment result, obtain the elements of exterior orientation of one camera in the mapping coordinate system of ground, then by the forward intersection method in photogrammetric, calculate 3 d space coordinate corresponding to observation station,

(12) sky three adjustment result that basis obtains and DEM or DSM carry out orthorectify to the polyphaser resultant image of step (6) splicing and make orthophotoquad, point on DEM or DSM is first projected on polyphaser resultant image, determine and fall into the corresponding district of which one camera raw video, DEM or DSM being pressed to the collinearity equation formula that step (8) sets up projects on this one camera raw video again, obtain the image coordinate of restoring on this one camera raw video, mathematics Conversion Relations tight between the one camera raw video of setting up according to step (4) again and polyphaser resultant image is determined the position on resultant image, polyphaser resultant image is carried out to sampled grey and give respective pixel on orthography.

It should be noted that, described one camera comprises single-lens battle array digital camera or its combination of various focal lengths.

Need to further illustrate, described multicamera system by two the above one cameras, by diverse location and different directions, placed and can be once on a large scale wide-angle obtain large format area array cameras or the panorama camera that space object image forms.

As a kind of preferred scheme, the object that described camera carrying platform comprises people's aircraft in the air, aerial unmanned vehicle, ground mobile vehicle, ground guide, ground moving or fixed support and pedestrian can be carried multicamera system; Described supporting positioning navigating device refers to GPS and IMU or both combination POS system.

Need to further illustrate, the one camera raw video in described step (2) is the raw video directly being obtained by one camera or after distortion correction and gray scale correction, can be directly used for carrying out from above black and white or the chromatic image that generation polyphaser resultant image is spliced in gray scale sampling.

It should be noted that, in described step (4), selected polyphaser resultant image splices required Method of Projection Change and can in certain accuracy rating, realize tight conversion mutually or mapping mutually between polyphaser resultant image and one camera raw video, comprising plane projection conversion, cylindrical surface projecting conversion, spherical projection converts and the table of comparisons is searched conversion.

It should be noted that, the polyphaser resultant image in described step (5) is directly or indirectly the raw video of one camera to be spliced to plane projection image, cylindrical surface projecting image and the spherical projection full-view image obtaining according to Method of Projection Change by large format area array cameras or panorama camera.

It should be noted that, on polyphaser resultant image in described step (5), each pixel can accurately transform on its corresponding one camera raw video and obtain the picpointed coordinate observed reading on one camera raw video by Method of Projection Change, and for the foundation of collinearity equation step (8) Suo Shu, the described foundation of error equation of step (9), the described forward intersection three-dimensional measuring of sky three adjustments of step (10) and step (11).

It should be noted that, two kinds of collinearity equation formulas that described step (8) is set up are respectively under aeroplane photography pattern and ground photography pattern, to process the general formula of polyphaser resultant image separately, picpointed coordinate in formula be by polyphaser resultant image through Method of Projection Change be transformed on one camera raw video or by one camera raw video on directly the collinearity equation formula of observation all set up, and can be according to the precision situation of the location navigation data of using in step (2) sky three adjustments with high precision GSP locator data, with sky three adjustments of high precision POS data and low precision or without sky three adjustments of location navigation data, can do self calibration adjustment according to result multicamera system is carried out to system calibration.

It should be noted that, three-dimensional measuring in described step (10) comprises the image measurement of double image stereo measurement and many baselines, during by ground point three dimensional space coordinate corresponding to the forward intersection calculated amount measuring point in photogrammetric, image coordinate used be with step (7) in for the adjustment point of empty three adjustments, through same treatment method, obtain, elements of exterior orientation used be by the elements of exterior orientation of the multicamera system obtaining after empty three adjustments of step (9) with corresponding one camera the orientation element in multicamera system by following formula in conjunction with after numerical value, wherein the calculating formula of elements of exterior orientation vertical element combination is:

$\left[\begin{array}{c}{X}_S\\ Y_S\\ Z_S\end{array}\right]=\left[\begin{array}{c}{X}_{\mathrm{GPSO}}+X_{\mathrm{GPS}}\\ Y_{\mathrm{GPSO}}+Y_{\mathrm{GPS}}\\ Z_{\mathrm{GPSO}}+Z_{\mathrm{GPS}}\end{array}\right]+R_{\mathrm{IMUO}}\·R_{\mathrm{IMU}}\·R_{\mathrm{virt}}(\left[\begin{array}{c}\mathrm{\Δ}{X}_{\mathrm{GPS}-\mathrm{virt}}\\ \mathrm{\Δ}{Y}_{\mathrm{GPS}-\mathrm{virt}}\\ \mathrm{\Δ}{Z}_{\mathrm{GPS}-\mathrm{virt}}\end{array}\right]+\left[\begin{array}{c}{X}_S^{\mathrm{sub}}\\ Y_S^{\mathrm{sub}}\\ Z_S^{\mathrm{sub}}\end{array}\right])---\left(4\right);$

The calculating formula of elements of exterior orientation angle element combination is:

R＝R _IMUO·R _IMU·R _virt·M _sub (5)。

Accompanying drawing explanation

Fig. 1 is the flow chart of data processing figure of the embodiment of the present invention;

Fig. 2 is the structural representation of the present invention's multicamera system in aeroplane photography pattern situation;

Fig. 3 is a kind of schematic diagram of the width polyphaser resultant image being spliced into by plane projection mapping mode;

Fig. 4 is a kind of schematic diagram of the width panorama polyphaser resultant image being spliced into by cylindrical surface projecting mapping mode;

Fig. 5 is a kind of schematic diagram of the width panorama polyphaser resultant image being spliced into by spherical projection mapping mode;

Reference numeral is explained:

210 represent multicamera system; 209 represent to form the one camera of multicamera system; 208 represent photographic, are here ground; 207 represent photo distance; 206 represent single-phase owner's distance; 205 represent the main distance of multicamera system; 204 represent one camera object lens center; 203 represent multicamera system coordinate origin, or the projection centre of polyphaser resultant image; 202 represent one camera raw video; 201 represent polyphaser resultant image; 212 represent the picture point of observing on polyphaser resultant image; 213 represent ground point corresponding to picture point of observing on polyphaser resultant image; 211 represent that the picture point back projection of observing on polyphaser resultant images is to the point on one camera image.

Embodiment

Embodiment below by aeroplane photography pattern, does further and illustrates technical solution of the present invention by reference to the accompanying drawings; It should be noted that, identical therewith to implementation process of the present invention under ground photography pattern.

As shown in Figure 1, the present invention is a kind of method that improves polyphaser resultant image photogrammetric accuracy, and described method is carried out according to the following steps:

Step 1, in 101, multicamera system 210 is carried out to calibration or the data that provides by 102Cong producer in obtain the elements of interior orientation, object lens distortion of each one camera 209 in multicamera system 210 and take the camera parameter 105 of definite elements of exterior orientation in the 203 polyphaser coordinate systems that are initial point;

Step 2, in 103 according to the requirement of measure the item with there being the target 208 in 104 pairs of measure the item locations of positioning navigating device such as the multicamera system 210 that carries on the platforms such as people's aircraft, aerial unmanned vehicle, ground mobile vehicle, ground guide, ground moving or fixed support or pedestrian and supporting GPS and IMU to photograph in the air, obtain raw video 202 and GPS and the IMU location navigation data 109 of all one cameras 209 in multicamera system 210; It should be noted that, the high-precision GPS of photogrammetric middle use and IMU location navigation data can reduce ground control point quantity, simplify empty three adjustments or not need empty three adjustments directly the aviation of obtaining or ground image to be positioned, raise the efficiency and Result Precision, save cost; High-precision GPS and IMU positioning navigating device are installed in suggestion in force.

Step 3 observes the three-dimensional coordinate obtaining in ground control point ground space three-dimensional system of coordinate evaluate for sky three adjustments 114 and Result Precision below in 109 to the ground control point in project location district according to the requirement of measure the item;

Step 4, in 107, according to the required Method of Projection Change of selected polyphaser resultant image 201 splicing of the type of multicamera system 210, and the camera parameter 105 obtaining according to step 1 is set up 201 tight image coordinate Conversion Relations of one camera raw video 202 and polyphaser resultant image.Selected Method of Projection Change can be realized 202 tight image coordinate of polyphaser resultant image 201 and one camera raw video and mutually changes or mutually shine upon in certain accuracy rating; As shown in Fig. 3～Fig. 5, comprise that plane projection conversion, cylindrical surface projecting convert, spherical projection converts and the table of comparisons is searched conversion, wherein 201 represent polyphaser resultant images; 301 represent that one camera raw video projects to the boundary line on resultant image; 302 represent that one camera raw video projects to the corresponding region on resultant image; 1～6 represents one camera.

In order further to understand the present invention, with a kind of simple Method of Projection Change under aeroplane photography pattern, the principle of 201 tight image coordinate Conversion Relations 107 of one camera raw video 202 and polyphaser resultant image is described below:

One camera raw video 202 is to the conversion on polyphaser resultant image 201, and transformation for mula is:

$\left\{\begin{array}{c}{x}_{\mathrm{virt}}=-f_{\mathrm{virt}}\frac{X_{\mathrm{virt}}}{Z_{\mathrm{virt}}}\\ y_{\mathrm{virt}}=-f_{\mathrm{virt}}\frac{Y_{\mathrm{virt}}}{Z_{\mathrm{virt}}}\end{array}\right.;$

Wherein, $\left[\begin{array}{c}{X}_{\mathrm{virt}}\\ Y_{\mathrm{virt}}\\ Z_{\mathrm{virt}}\end{array}\right]=M_{\mathrm{sub}}\left[\begin{array}{c}{x}_{\mathrm{sub}}-x_{\mathrm{sub}}^0+\mathrm{\Δ}{x}_{\mathrm{sub}}\\ y_{\mathrm{sub}}-y_{\mathrm{sub}}^0+\mathrm{\Δ}{y}_{\mathrm{sub}}\\ -f_{\mathrm{sub}}\end{array}\right];$

M _subfor one camera 209 definite deflection element (roll angle ω in multicamera system 210 coordinate systems _sub, the angle of pitch

with position angle κ _sub) rotation matrix that forms, expression is:

Wherein, x _sub, y _subimage coordinate for the picture point 211 on one camera raw video 202;

principal point position for one camera 209; f _sub main distance 206 for one camera 209; Δ x _sub, Δ y _subfor x _sub, y _subin the systematic error compensation that caused by one camera raw video 202 distortion and object lens distortion etc.; x _virt, y _virtfor transforming to picture point 212 image coordinate on polyphaser resultant image 201;

F _virt main distance 205 for multicamera system 210.

Polyphaser resultant image 201 is to the conversion on one camera raw video 202, and transformation for mula is:

$\left\{\begin{array}{c}{x}_{\mathrm{sub}}=-f_{\mathrm{sub}}\frac{X_{\mathrm{sub}}}{Z_{\mathrm{sub}}}+x_{\mathrm{sub}}^0-\mathrm{\Δ}{x}_{\mathrm{sub}}\\ y_{\mathrm{sub}}=-f_{\mathrm{sub}}\frac{Y_{\mathrm{sub}}}{Z_{\mathrm{sub}}}+y_{\mathrm{sub}}^0-\mathrm{\Δ}{y}_{\mathrm{sub}}\end{array}\right.$

Wherein, $\left[\begin{array}{c}{X}_{\mathrm{sub}}\\ Y_{\mathrm{sub}}\\ Z_{\mathrm{sub}}\end{array}\right]=M_{\mathrm{sub}}^{-1}\left[\begin{array}{c}{x}_{\mathrm{virt}}\\ y_{\mathrm{virt}}\\ -f_{\mathrm{virt}}\end{array}\right].$

Realizing the method that in technical solution of the present invention, between polyphaser resultant image and one camera raw video, image coordinate is changed mutually has multiple, above-mentioned just wherein a kind of, object is in order to set up the collinearity equation formula of the tight Photogrammetric Processing of polyphaser resultant image, therefore, should not be construed as limitation of the present invention.From the simple Conversion Relations of 202 of the polyphaser resultant image 201 of above-mentioned aeroplane photography pattern and one camera raw videos, can find out that both are reversible in certain accuracy rating.The technical scheme of implementing according to this principle all should belong to the scope of the claims in the present invention.

Step 5, each in 108, step 2 being obtained taken the photograph 201 tight image coordinate Conversion Relations 107 of one camera raw video 202 that station 203 one camera raw video 202 sets up by step 4 and polyphaser resultant image and spliced, obtain polyphaser resultant image 201 and elements of interior orientation thereof that each takes the photograph station, and 201 mark the corresponding region 302 of one camera raw video on polyphaser resultant image;

Step 6, the one camera parameter 105 of utilizing location navigation data 109 that step 2 obtains, polyphaser resultant image 201 that step 5 obtains and elements of interior orientation thereof and step 1 to obtain according to the requirement of measure the item in 110 is set up the photogrammetric data that empty three adjustments, three-dimensional measuring and orthorectify use and is processed engineering;

Step 7, the polyphaser resultant image 201 that the photogrammetric data processing engineering of setting up according to step 6 in 111 obtains step 5 is encrypted the automatic and manual observation of the image coordinate of a picture point 212, the image coordinate of the corresponding picture point 212 of the ground control point that step 3 is obtained on polyphaser resultant image 201 is carried out manual observation, and the corresponding region 302 of one camera raw video 202 marking on the polyphaser resultant image 201 that obtains of 201 tight image coordinate Conversion Relations 107 of the one camera raw video 202 of in 112, the image coordinate of this 2 class point 212 on polyphaser resultant image 201 being set up by step 4 and polyphaser resultant image and step 5 transforms on one camera raw video 202, obtain the image coordinate of the picture point 211 on one camera raw video 202 for sky three compensating computations in the foundation and 114 of 113 collinearity equation formulas,

Step 8, in 113, according to the photograph mode of multicamera system 210, be aeroplane photography or ground photography, here select aeroplane photography pattern, on the one camera raw video that 109 pairs of steps 7 of location navigation data that integrating step 2 obtains obtain, 202 picture point 211, this picture point 211 place one camera raw video projection centre 204 ground point 213 corresponding with this picture point are set up the collinearity equation formula that the tight Photogrammetric Processing of polyphaser resultant image 201 is used, and the collinearity equation formula under aeroplane photography pattern is:

$\left[\begin{array}{c}{x}_{\mathrm{sub}}-x_{\mathrm{sub}}^0+\mathrm{\Δ}{x}_{\mathrm{sub}}\\ y_{\mathrm{sub}}-y_{\mathrm{sub}}^0+\mathrm{\Δ}{y}_{\mathrm{sub}}\\ -f_{\mathrm{sub}}\end{array}\right]=\frac{1}{\mathrm{\λ}}{M}_{\mathrm{sub}}^{-1}(R_{\mathrm{virt}}^{-1}\·R_{\mathrm{IMU}}^{-1}\left[\begin{array}{c}X-X_{\mathrm{GPS}}\\ Y-Y_{\mathrm{GPS}}\\ Z-Z_{\mathrm{GPS}}\end{array}\right]-\left[\begin{array}{c}\mathrm{\Δ}{X}_{\mathrm{GPS}-\mathrm{virt}}+X_S^{\mathrm{sub}}\\ \mathrm{\Δ}{Y}_{\mathrm{GPS}-\mathrm{virt}}+Y_S^{\mathrm{sub}}\\ \mathrm{\Δ}{Z}_{\mathrm{GPS}-\mathrm{virt}}+Z_S^{\mathrm{sub}}\end{array}\right])---\left(1\right)$

Wherein, X, Y, Z is the three-dimensional coordinate of millet cake 213 in object space mapping coordinate system accordingly, X _gPS, Y _gPS, Z _gPSfor the three-dimensional coordinate that positioning navigating device in step 2 is observed in object space mapping coordinate system when multicamera system exposes, R _iMUfor the rotation matrix that the attitude angle (roll angle ω, angle of pitch φ and position angle κ) of observation forms when multicamera system exposes in object space mapping coordinate system of positioning navigating device in step 2, R _virtfor inconsistent amount (the roll angle ω of positioning navigating device coordinate system direction in multicamera system coordinate system and step 2 _virt, the angle of pitch

with position angle κ _virt) rotation matrix that forms, M _subfor one camera definite deflection element (roll angle ω in multicamera system coordinate system _sub, the angle of pitch

with position angle κ _sub) rotation matrix that forms, Δ X _gPS-virt, Δ Y _gPS-virt, Δ Z _gPS-virtfor gps antenna center definite relative position in polyphaser coordinate system,

for one camera object lens center 204 definite relative position in polyphaser coordinate system, x _sub, y _subpicture point 211 image coordinate on the one camera raw video 202 obtaining for step 7,

for principal point 204 positions of one camera 209, f _subfor the main distance 206 of one camera, Δ x _sub, Δ y _subfor x _sub, y _subin the systematic error compensation that caused by one camera deformation of image and object lens distortion etc.;

Step 9, the collinearity equation formula of in 114, step 8 being set up is carried out linearization and is obtained following error equation:

$\left\{\begin{array}{cccccccc}{V}_P=C_{\mathrm{GPSIMU}}{X}_{\mathrm{GPSIMU}}+C_G{X}_G+C_A{X}_A+C_{\mathrm{GPS}-\mathrm{virt}}{X}_{\mathrm{GPS}-\mathrm{virt}}+C_{\mathrm{IMU}-\mathrm{virt}}{X}_{\mathrm{IMU}-\mathrm{virt}}+C_{\mathrm{virt}-\mathrm{sub}}{X}_{\mathrm{virt}-\mathrm{sub}}& & & & & & -L_P& P_P\\ V_G=& E_G{X}_G& & & & & -L_G& P_G\\ V_A=& & E_A{X}_A& & & & -L_A& P_A\\ V_{\mathrm{GPSIMU}}=E_{\mathrm{GPSIMU}}{X}_{\mathrm{GPSIMU}}& & & & & & -L_{\mathrm{GPSIMU}}& P_{\mathrm{GPSIMU}}\\ V_{\mathrm{GPS}-\mathrm{virt}}=& & & E_{\mathrm{GPS}-\mathrm{virt}}{X}_{\mathrm{GPS}-\mathrm{virt}}& & & -L_{\mathrm{GPS}-\mathrm{virt}}& P_{\mathrm{GPS}-\mathrm{virt}}\\ V_{\mathrm{IMU}-\mathrm{virt}}=& & & & E_{\mathrm{IMU}-\mathrm{virt}}{X}_{\mathrm{IMU}-\mathrm{virt}}& & -L_{\mathrm{IMU}-\mathrm{virt}}& P_{\mathrm{IMU}-\mathrm{virt}}\\ V_{\mathrm{virt}-\mathrm{sub}}=& & & & & E_{\mathrm{virt}-\mathrm{sub}}{X}_{\mathrm{virt}-\mathrm{sub}}& -L_{\mathrm{virt}-\mathrm{sub}}& P_{\mathrm{virt}-\mathrm{sub}}\end{array}\right.---\left(3\right)$

Wherein, VP, VG, VA, VGPS-IMU, VGPS-virt, VIMU-virt and Vvirt-sub be respectively one camera raw video coordinate, ground point object space three-dimensional coordinate, self calibration additional parameter, GPS/IMU location navigation data, gps antenna center to polyphaser resultant image projection centre side-play amount, polyphaser coordinate system and IMU coordinate system is inconsistent and one camera at polyphaser

The observed reading of the orientation element in coordinate system corrects vector, XG, XA, XGPS-IMU, XGPS-virt, XIMU-virt and Xvirt-sub are respectively ground point object space three-dimensional coordinate, self calibration additional parameter, GPS/IMU location navigation data, gps antenna center is to polyphaser resultant image projection centre side-play amount, polyphaser coordinate system and IMU coordinate system are inconsistent, and the orientation element of one camera in polyphaser coordinate system is as the correction vector of adjustment unknown number, CG, CA, CGPS-IMU, CGPS-virt, CIMU-virt and Cvirt-sub are respectively XG, XA, XGPS-IMU, XGPS-virt, the matrix of coefficients that XIMU-virt and Xvirt-sub are corresponding, EG, EA, EGPS-IMU, EGPS-virt, EIMU-virt and Evirt-sub are respectively XG, XA, XGPS-IMU, XGPS-virt, the unit coefficient matrix that XIMU-virt and Xvirt-sub are corresponding, LP, LG, LA, LGPS-IMU, LGPS-virt, LIMU-virt and Lvirt-sub are respectively VP, VG, VA, VGPS-IMU, VGPS-virt, the constant vector that VIMU-virt and Vvirt-sub are corresponding, PP, PG, PA, PGPS-IMU, PGPS-virt, PIMU-virt and Pvirt-sub are respectively one camera raw video coordinate, ground point object space three-dimensional coordinate, self calibration additional parameter, GPS/IMU location navigation data, gps antenna center is to polyphaser resultant image projection centre side-play amount, polyphaser coordinate system and IMU coordinate system are inconsistent, and the weight matrix corresponding to observed reading of the orientation element of one camera in polyphaser coordinate system, the precision that reflects these observed readings.Collinearity equation formula under ground photography pattern also can obtain the error equation of same form and carry out sky three compensating computations below after linearization;

Step 10, the error equation of simultaneously step 9 being listed in 114 is according to the method for existing photogrammetric hollow three adjustments, by the least square method methodization of carrying out, form normal equation, by interative computation, try to achieve the unknown number in error equation, when resolving sky three adjustment processing that complete polyphaser resultant image while meeting predefined precision conditions, obtain the end product of empty three adjustments;

Step 11, sky three adjustment result that obtain according to step 10 in 115 are carried out three-dimensional measuring to polyphaser resultant image, the picpointed coordinate of observing on polyphaser resultant image is transformed on one camera raw video by the processing mode described in step 7, obtain the image coordinate on one camera raw video, the elements of exterior orientation of the multicamera system in sky three adjustment result that again step 10 obtained is two formula combinations below the use of the orientation element in multicamera system with corresponding one camera, obtain the elements of exterior orientation of one camera in the mapping coordinate system of ground, then by the forward intersection method in photogrammetric, calculate 3 d space coordinate corresponding to observation station, one camera elements of exterior orientation vertical element is in conjunction with calculating formula:

$\left[\begin{array}{c}{X}_S\\ Y_S\\ Z_S\end{array}\right]=\left[\begin{array}{c}{X}_{\mathrm{GPS}}\\ Y_{\mathrm{GPS}}\\ Z_{\mathrm{GPS}}\end{array}\right]+R_{\mathrm{IMU}}\·R_{\mathrm{virt}}(\left[\begin{array}{c}\mathrm{\Δ}{X}_{\mathrm{GPS}-\mathrm{vist}}\\ \mathrm{\Δ}{Y}_{\mathrm{GPS}-\mathrm{vist}}\\ \mathrm{\Δ}{Z}_{\mathrm{GPS}-\mathrm{vist}}\end{array}\right]+\left[\begin{array}{c}{X}_S^{\mathrm{sub}}\\ Y_S^{\mathrm{sub}}\\ Z_S^{\mathrm{sub}}\end{array}\right])---\left(4\right)$

One camera elements of exterior orientation angle element is in conjunction with calculating formula:

R＝R _IMU·R _virt·M _sub (5)。

Step 12, sky three adjustment result that obtain according to step 10 in 115 and DEM or DSM carry out orthorectify to the polyphaser resultant image of step 6 splicing and make orthophotoquad, point on DEM or DSM is first projected on polyphaser resultant image, determine and fall into the corresponding district of which one camera raw video, DEM or DSM being pressed to the collinearity equation formula that step 8 sets up projects on this one camera raw video again, obtain the image coordinate of restoring on this one camera raw video, mathematics Conversion Relations tight between the one camera raw video of setting up according to step 4 again and polyphaser resultant image is determined the position on resultant image, carrying out polyphaser resultant image carries out sampled grey and gives on orthography,

Step 13, sky three adjustment result that obtain according to step 10 in 115 are by make method data texturing from polyphaser image the automatically collection required to three-dimensional modeling of orthophotoquad in step 12 of polyphaser resultant image.

In order to further describe effect of the present invention, according to the step described in technique scheme, the multicamera system that adopts an EOS5D MarkII of Ge Youliangtai Canon to form is installed to unmanned airship, and with differential GPS, 300 meters of flying heights, implement on the spot the outdoor of ground resolution 8cm, adopts classic method and method of the present invention described in document to process polyphaser resultant image under equal test condition, and table 1 has been listed sky three test findings of these two kinds of methods.

Empty three results of implementation of table 1 the present invention and the empty three result precision of tradition

Same according to the step described in technique scheme, at an indoor multicamera system being formed by 5 EOS5D Mark II of Canon, be suspended on a moving track that overhead 10m is high and photographed in ground, be not with differential GPS, ground resolution is 2mm, also adopt classic method and method of the present invention described in document to process polyphaser resultant image, table 2 has been listed the test result of these two kinds of methods.

Empty three results of implementation of the indoor the present invention of table 2 and the empty three result precision of tradition

The result obtaining from table 1 and two kinds of test figures of table 2 can be found out, the result obtaining under the result that method of the present invention is done under outdoor four kinds of test conditions and indoor two kinds of test conditions, the precision of the result that precision all obtains under the same conditions higher than classic method, has verified the actual effect of the method for the invention.

It should be noted that, in table 1 and table 2, Sigma represents error in the weight unit of empty three adjustments, and unit is pixel.XYZ is the middle error of 3 d space coordinate, and unit is rice.

For a person skilled in the art, can make other various corresponding changes and distortion according to technical scheme described above and design, and these all changes and distortion all should belong to the protection domain of the claims in the present invention within.

Claims (10)

1. a method that improves polyphaser resultant image photogrammetric accuracy, is characterized in that, said method comprising the steps of:

(1) multicamera system is carried out to calibration obtains the distortion of each one camera elements of interior orientation in multicamera system, object lens and the camera parameter of definite elements of exterior orientation in polyphaser coordinate system;

(2) described multicamera system is installed on camera carrying platform according to the requirement of measure the item and supporting positioning navigating device is photographed to the target in described measure the item location, and is obtained raw video and the location navigation data of all one cameras in described multicamera system;

(3) according to the requirement of measure the item, the ground control point in project location district is observed; (4) according to the selected polyphaser resultant image of the type of described multicamera system, splice required Method of Projection Change, and the camera parameter obtaining according to step (1) is set up image coordinate Conversion Relations tight between one camera raw video and polyphaser resultant image;

(5) each one camera raw video step (2) being obtained, then the image coordinate Conversion Relations obtaining by step (4) splices, obtain each polyphaser resultant image and elements of interior orientation thereof, and on polyphaser resultant image, mark the corresponding region of one camera raw video;

(6), according to the requirement of measure the item, the one camera parameter of utilizing location navigation data that step (2) obtains, polyphaser resultant image that step (5) obtains and elements of interior orientation thereof and step (1) to obtain is set up the photogrammetric data that empty three adjustments, three-dimensional measuring and orthorectify use and is processed engineering;

(7) photogrammetric data of setting up according to step (6) process polyphaser resultant image that engineering obtains step (5) be encrypted a picpointed coordinate (what is manual observation with manual observation automatically?), the corresponding picpointed coordinate of the ground control point that step (3) is obtained on polyphaser resultant image carries out manual observation, and the corresponding region of one camera raw video marking on the polyphaser resultant image that the image coordinate of this 2 class point on polyphaser resultant image obtained by the image coordinate Conversion Relations of step (4) foundation and step (5) transforms on one camera raw video, obtain the image coordinate on one camera raw video,

(8) picture point on the one camera raw video that the location navigation data that integrating step (2) obtains obtain step (7), the described picture point place one camera raw video projection centre ground point corresponding with it are set up respectively corresponding collinearity equation formula according to the pattern of aeroplane photography or ground photography, and the collinearity equation formula under aeroplane photography pattern is:

$\left[\begin{array}{c}X\\ Y\\ Z\end{array}\right]=\left[\begin{array}{c}{X}_{\mathrm{GPSO}}+X_{\mathrm{GPS}}\\ Y_{\mathrm{GPSO}}+Y_{\mathrm{GPS}}\\ Z_{\mathrm{GPSO}}+Z_{\mathrm{GPS}}\end{array}\right]+R_{\mathrm{IMUO}}\·R_{\mathrm{IMU}}\·R_{\mathrm{virt}}(\left[\begin{array}{c}\mathrm{\Δ}{X}_{\mathrm{GPS}-\mathrm{virt}}\\ \mathrm{\Δ}{Y}_{\mathrm{GPS}-\mathrm{virt}}\\ \mathrm{\Δ}{Z}_{\mathrm{GPS}-\mathrm{virt}}\end{array}\right]+\left[\begin{array}{c}{X}_S^{\mathrm{sub}}\\ Y_S^{\mathrm{sub}}\\ Z_S^{\mathrm{sub}}\end{array}\right]+\mathrm{\λ}{M}_{\mathrm{sub}}\left[\begin{array}{c}{x}_{\mathrm{sub}}-x_{\mathrm{sub}}^0+\mathrm{\Δ}{x}_{\mathrm{sub}}\\ y_{\mathrm{sub}}-y_{\mathrm{sub}}^0+\mathrm{\Δ}{y}_{\mathrm{sub}}\\ -f_{\mathrm{sub}}\end{array}\right])---\left(1\right)$

Wherein, X, Y, Z is the three-dimensional coordinate of millet cake in object space mapping coordinate system accordingly, X _gPS, Y _gPS, Z _gPSfor the three-dimensional coordinate of taking the photograph station that positioning navigating device in step (2) is obtained in object space mapping coordinate system, X _gPSO, Y _gPSO, Z _gPSOfor X _gPS, Y _gPS, Z _gPSthe value of starting in object space mapping coordinate system, R _iMUfor the attitude angle of taking the photograph station that positioning navigating device in step 2 is obtained in object space mapping coordinate system, the rotation matrix being formed by roll angle ω, angle of pitch φ and position angle κ, R _iMUOfor the attitude angle of taking the photograph station value of starting in object space mapping coordinate system, the rotation matrix being formed by roll angle ω 0, angle of pitch φ 0 and position angle κ 0, R _virtfor the inconsistent amount of positioning navigating device coordinate system direction in multicamera system coordinate system and step (1), the rotation matrix being formed by roll angle ω virt, angle of pitch φ virt and position angle κ virt, M _subfor one camera definite deflection element in multicamera system coordinate system, the rotation matrix being formed by roll angle ω sub, angle of pitch φ sub and position angle κ sub, Δ X _gPS-virt, Δ Y _gPS-virt, Δ Z _gPS-virtfor gps antenna center definite relative position in polyphaser coordinate system,

for one camera object lens center definite relative position in polyphaser coordinate system, x _sub, y _subpicture point image coordinate on the one camera raw video obtaining for step (7),

for the principal point position of one camera, f _subfor the main distance of one camera, Δ x _sub, Δ y _subfor x _sub, y _subin the systematic error compensation that caused by one camera deformation of image and object lens distortion etc.;

Collinearity equation formula under ground photography pattern is:

$\left[\begin{array}{c}X\\ Y\\ Z\end{array}\right]=\left[\begin{array}{c}{X}_{\mathrm{GPSO}}+X_{\mathrm{GPS}}\\ Y_{\mathrm{GPSO}}+Y_{\mathrm{GPS}}\\ Z_{\mathrm{GPSO}}+Z_{\mathrm{GPS}}\end{array}\right]+R_{\mathrm{IMUO}}\·R_{\mathrm{IMU}}\·R_{\mathrm{virt}}(\left[\begin{array}{c}\mathrm{\Δ}{X}_{\mathrm{GPS}-\mathrm{virt}}\\ \mathrm{\Δ}{Y}_{\mathrm{GPS}-\mathrm{virt}}\\ \mathrm{\Δ}{Z}_{\mathrm{GPS}-\mathrm{virt}}\end{array}\right]+\left[\begin{array}{c}{X}_S^{\mathrm{sub}}\\ Y_S^{\mathrm{sub}}\\ Z_S^{\mathrm{sub}}\end{array}\right]+\mathrm{\λ}{M}_{\mathrm{sub}}\left[\begin{array}{c}{x}_{\mathrm{sub}}-x_{\mathrm{sub}}^0+\mathrm{\Δ}{x}_{\mathrm{sub}}\\ f_{\mathrm{sub}}\\ y_{\mathrm{sub}}-y_{\mathrm{sub}}^0+\mathrm{\Δ}{y}_{\mathrm{sub}}\end{array}\right])---\left(2\right)$

Wherein, R _iMUO, R _iMU, R _virtand M _subbe respectively the rotation matrix usually forming according to the angle unit of ground photograph mode definition, the implication of other symbol representative is identical with formula (1);

(9) the aeroplane photography pattern of step (8) being set up and the collinearity equation formula under ground photography pattern are all carried out linearization by same mode and are obtained following identic error equation:

$\left\{\begin{array}{cccccccc}{V}_P=C_{\mathrm{GPSIMU}}{X}_{\mathrm{GPSIMU}}+C_G{X}_G+C_A{X}_A+C_{\mathrm{GPS}-\mathrm{virt}}{X}_{\mathrm{GPS}-\mathrm{virt}}+C_{\mathrm{IMU}-\mathrm{virt}}{X}_{\mathrm{IMU}-\mathrm{virt}}+C_{\mathrm{virt}-\mathrm{sub}}{X}_{\mathrm{virt}-\mathrm{sub}}& & & & & & -L_P& P_P\\ V_G=& E_G{X}_G& & & & & -L_G& P_G\\ V_A=& & E_A{X}_A& & & & -L_A& P_A\\ V_{\mathrm{GPSIMU}}=E_{\mathrm{GPSIMU}}{X}_{\mathrm{GPSIMU}}& & & & & & -L_{\mathrm{GPSIMU}}& P_{\mathrm{GPSIMU}}\\ V_{\mathrm{GPS}-\mathrm{virt}}=& & & E_{\mathrm{GPS}-\mathrm{virt}}{X}_{\mathrm{GPS}-\mathrm{virt}}& & & -L_{\mathrm{GPS}-\mathrm{virt}}& P_{\mathrm{GPS}-\mathrm{virt}}\\ V_{\mathrm{IMU}-\mathrm{virt}}=& & & & E_{\mathrm{IMU}-\mathrm{virt}}{X}_{\mathrm{IMU}-\mathrm{virt}}& & -L_{\mathrm{IMU}-\mathrm{virt}}& P_{\mathrm{IMU}-\mathrm{virt}}\\ V_{\mathrm{virt}-\mathrm{sub}}=& & & & & E_{\mathrm{virt}-\mathrm{sub}}{X}_{\mathrm{virt}-\mathrm{sub}}& -L_{\mathrm{virt}-\mathrm{sub}}& P_{\mathrm{virt}-\mathrm{sub}}\end{array}\right.---\left(3\right)$

Wherein, VP, VG, VA, VGPS-IMU, VGPS-virt, VIMU-virt and Vvirt-sub are respectively one camera raw video coordinate, ground point object space three-dimensional coordinate, self calibration additional parameter, GPS/IMU location navigation data, gps antenna center is to polyphaser resultant image projection centre side-play amount, polyphaser coordinate system and IMU coordinate system are inconsistent, and the observed reading of the orientation element of one camera in polyphaser coordinate system corrects vector, XG, XA, XGPS-IMU, XGPS-virt, XIMU-virt and Xvirt-sub are respectively ground point object space three-dimensional coordinate, self calibration additional parameter, GPS/IMU location navigation data, gps antenna center is to polyphaser resultant image projection centre side-play amount, polyphaser coordinate system and IMU coordinate system are inconsistent, and the orientation element of one camera in polyphaser coordinate system is as the correction vector of adjustment unknown number, CG, CA, CGPS-IMU, CGPS-virt, CIMU-virt and Cvirt-sub are respectively XG, XA, XGPS-IMU, XGPS-virt, the matrix of coefficients that XIMU-virt and Xvirt-sub are corresponding, EG, EA, EGPS-IMU, EGPS-virt, EIMU-virt and Evirt-sub are respectively XG, XA, XGPS-IMU, XGPS-virt, the unit coefficient matrix that XIMU-virt and Xvirt-sub are corresponding, LP, LG, LA, LGPS-IMU, LGPS-virt, LIMU-virt and Lvirt-sub are respectively VP, VG, VA, VGPS-IMU, VGPS-virt, VIMU-virt and constant vector corresponding to Vvi rt-sub, PP, PG, PA, PGPS-IMU, PGPS-virt, PIMU-virt and Pvirt-sub are respectively one camera raw video coordinate, ground point object space three-dimensional coordinate, self calibration additional parameter, GPS/IMU location navigation data, gps antenna center is to polyphaser resultant image projection centre side-play amount, polyphaser coordinate system and IMU coordinate system are inconsistent, and the weight matrix corresponding to observed reading of the orientation element of one camera in polyphaser coordinate system, the precision that reflects these observed readings,

(10) according to the method for photogrammetric hollow three adjustments, by the error equation that least square method is set up step (9) methodization of carrying out, form normal equation, by interative computation, try to achieve the unknown number in error equation, when resolving sky three adjustment processing that complete polyphaser resultant image while meeting predefined precision conditions, obtain the end product of empty three adjustments;

(11) according to sky three adjustment result that obtain, polyphaser resultant image is carried out to three-dimensional measuring, the picpointed coordinate of observing on polyphaser resultant image is transformed on one camera raw video by processing mode step (7) Suo Shu, obtain the image coordinate on one camera raw video, again by the orientation element combination in multicamera system with corresponding one camera of the elements of exterior orientation of the multicamera system in the sky obtaining three adjustment result, obtain the elements of exterior orientation of one camera in the mapping coordinate system of ground, then by the forward intersection method in photogrammetric, calculate 3 d space coordinate corresponding to observation station,

(12) sky three adjustment result that basis obtains and DEM or DSM carry out orthorectify to the polyphaser resultant image of step (6) splicing and make orthophotoquad, point on DEM or DSM is first projected on polyphaser resultant image, determine and fall into the corresponding district of which one camera raw video, DEM or DSM being pressed to the collinearity equation formula that step (8) sets up projects on this one camera raw video again, obtain the image coordinate of restoring on this one camera raw video, mathematics Conversion Relations tight between the one camera raw video of setting up according to step (4) again and polyphaser resultant image is determined the position on resultant image, polyphaser resultant image is carried out to sampled grey and give respective pixel on orthography.

2. raising polyphaser resultant image photogrammetric accuracy method according to claim 1, is characterized in that, described one camera comprises single-lens battle array digital camera or its combination of various focal lengths.

3. raising polyphaser resultant image photogrammetric accuracy method according to claim 1 and 2, it is characterized in that, described multicamera system by two the above one cameras, by diverse location and different directions, placed and can be once on a large scale wide-angle obtain large format area array cameras or the panorama camera that space object image forms.

4. raising polyphaser resultant image photogrammetric accuracy method according to claim 1, it is characterized in that the object that described camera carrying platform comprises people's aircraft in the air, aerial unmanned vehicle, ground mobile vehicle, ground guide, ground moving or fixed support and pedestrian can be carried multicamera system; Described supporting positioning navigating device refers to GPS and IMU or both combination POS system.

5. raising polyphaser resultant image photogrammetric accuracy method according to claim 1 and 2, it is characterized in that, the one camera raw video in described step (2) is the raw video directly being obtained by one camera or after distortion correction and gray scale correction, can be directly used for carrying out from above black and white or the chromatic image that generation polyphaser resultant image is spliced in gray scale sampling.

6. raising polyphaser resultant image photogrammetric accuracy method according to claim 1, it is characterized in that, in described step (4), selected polyphaser resultant image splices required Method of Projection Change and can in certain accuracy rating, realize tight conversion mutually or mapping mutually between polyphaser resultant image and one camera raw video, comprising plane projection conversion, cylindrical surface projecting conversion, spherical projection converts and the table of comparisons is searched conversion.

7. raising polyphaser resultant image photogrammetric accuracy method according to claim 1, it is characterized in that, the polyphaser resultant image in described step (5) is directly or indirectly the raw video of one camera to be spliced to plane projection image, cylindrical surface projecting image and the spherical projection full-view image obtaining according to Method of Projection Change by large format area array cameras or panorama camera.

8. raising polyphaser resultant image photogrammetric accuracy method according to claim 1, it is characterized in that, on polyphaser resultant image in described step (5), each pixel can accurately transform on its corresponding one camera raw video and obtain the picpointed coordinate observed reading on one camera raw video by Method of Projection Change, and for the foundation of collinearity equation step (8) Suo Shu, the described foundation of error equation of step (9), the described forward intersection three-dimensional measuring of sky three adjustments of step (10) and step (11).

9. raising polyphaser resultant image photogrammetric accuracy method according to claim 1, it is characterized in that, two kinds of collinearity equation formulas that described step (8) is set up are respectively under aeroplane photography pattern and ground photography pattern, to process the general formula of polyphaser resultant image separately, picpointed coordinate in formula be by polyphaser resultant image through Method of Projection Change be transformed on one camera raw video or by one camera raw video on directly the collinearity equation formula of observation all set up, and can be according to the precision situation of the location navigation data of using in step (2) sky three adjustments with high precision GSP locator data, with sky three adjustments of high precision POS data and low precision or without sky three adjustments of location navigation data, can do self calibration adjustment according to result multicamera system is carried out to system calibration.

10. raising polyphaser resultant image photogrammetric accuracy method according to claim 1, it is characterized in that, three-dimensional measuring in described step (10) comprises the image measurement of double image stereo measurement and many baselines, during by ground point three dimensional space coordinate corresponding to the forward intersection calculated amount measuring point in photogrammetric, image coordinate used be with step (7) in for the adjustment point of empty three adjustments, through same treatment method, obtain, elements of exterior orientation used be by the elements of exterior orientation of the multicamera system obtaining after empty three adjustments of step (9) with corresponding one camera the orientation element in multicamera system by following formula in conjunction with after numerical value, wherein the calculating formula of elements of exterior orientation vertical element combination is:

$\left[\begin{array}{c}{X}_S\\ Y_S\\ Z_S\end{array}\right]=\left[\begin{array}{c}{X}_{\mathrm{GPSO}}+X_{\mathrm{GPS}}\\ Y_{\mathrm{GPSO}}+Y_{\mathrm{GPS}}\\ Z_{\mathrm{GPSO}}+Z_{\mathrm{GPS}}\end{array}\right]+R_{\mathrm{IMUO}}\·R_{\mathrm{IMU}}\·R_{\mathrm{virt}}(\left[\begin{array}{c}\mathrm{\Δ}{X}_{\mathrm{GPS}-\mathrm{virt}}\\ \mathrm{\Δ}{Y}_{\mathrm{GPS}-\mathrm{virt}}\\ \mathrm{\Δ}{Z}_{\mathrm{GPS}-\mathrm{virt}}\end{array}\right]+\left[\begin{array}{c}{X}_S^{\mathrm{sub}}\\ Y_S^{\mathrm{sub}}\\ Z_S^{\mathrm{sub}}\end{array}\right])---\left(4\right);$

The calculating formula of elements of exterior orientation angle element combination is:

R＝R _IMUO·R _IMU·R _virt·M _sub(5)。

Images